Dynamic SoC Reconfiguration Without System Reset
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current System on a Chip (SoC) technologies are limited by their inability to dynamically reconfigure operational configurations without a complete reboot, restricting flexibility and efficiency in managing multiple independent operating environments.
Innovation Solution
The implementation of a dynamically reconfigurable SoC (DRSOC) that allows for changes in operational configurations without rebooting, utilizing a microsystem manager, resource manager, and blockchain-based collaborative verification to reallocate IP blocks and manage microsystems dynamically, ensuring secure and trusted boot processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional SoC configuration methods are used, then system stability is maintained through fixed configurations, but system flexibility and adaptability deteriorate due to inability to dynamically reconfigure
Solution Approach 1:
The SoC is divided into multiple independent IP blocks that can be individually allocated and reconfigured. Each IP block represents a separable functional unit that can be dynamically assigned to different microsystems, enabling flexible reconfiguration without affecting the entire system.
Solution Approach 2:
The system implements dynamic reconfiguration capabilities where IP block allocations can be changed at runtime without complete system reboot. The microsystem manager and resource manager enable real-time adjustments to configuration based on operational needs, transitioning from static to dynamic configuration management.
2Productivity
If dynamic reconfiguration without reboot is implemented, then system efficiency and productivity improve, but verification complexity and security risks increase
Solution Approach 1:
The system implements a feedback mechanism where the microsystem manager monitors configuration states and operational conditions, then adjusts IP block allocations accordingly. This closed-loop approach ensures that reconfiguration decisions are based on actual system state, improving both efficiency and reliability.
Solution Approach 2:
The resource manager acts as an intermediary between the microsystem manager and IP blocks, coordinating allocation decisions and ensuring proper verification. This intermediary layer manages the complexity of verification by centralizing control and coordination of reconfiguration operations.
3Quantity of substance
If multiple microsystems share IP blocks dynamically, then resource utilization improves, but conflict detection and measurement difficulty increase
Solution Approach 1:
The system uses standardized interfaces and protocols for IP block allocation and verification across all microsystems. This homogenization of interaction patterns simplifies the detection and measurement of configuration states, making verification more manageable despite dynamic sharing.
Solution Approach 2:
The verification mechanism combines multiple checking layers (resource manager verification, microsystem manager coordination, and operational monitoring) to create a composite verification system that can reliably detect configuration states despite complex dynamic sharing patterns.
Data Source
AI summary
A method dynamically reconfigures a system on a chip (SOC) comprising multiple semiconductor intellectual property (IP) blocks. The method comprises, when booting a data processing system (DPS) comprising the SOC, automatically allocating different IP blocks to multiple different microsystems within the DPS, based on a static partitioning policy (SPP). The method also comprises, after booting the DPS, determining that reallocation of at least one of the IP blocks is desired, based on (a) monitored conditions of at least one of the microsystems and (b) a dynamic partitioning policy (DPP). The method also comprises, in response to determining that reallocation of at least one of the IP blocks is desired, automatically reallocating at least one of the IP blocks from one of the microsystems to another of the microsystems without resetting at least one of the microsystems. Other embodiments are described and claimed.